Spaces:

OpenRAL
/

README

Running

App Files Files Community

AdrianLlopart commited on Jun 20

Commit

3af10a9

verified ·

1 Parent(s): 0413a39

Convert label–definition em-dashes to colons/periods (impeccable: em-dash overuse)

Browse files

Files changed (1) hide show

index.html +21 -21

index.html CHANGED Viewed

@@ -3,7 +3,7 @@
 <head>
   <meta charset="UTF-8" />
   <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-  <title>OpenRAL — Runtime for VLA Robot Agents</title>
   <style>
     *, *::before, *::after { box-sizing: border-box; margin: 0; padding: 0; }
@@ -227,10 +227,10 @@
     <p>A VLA model alone can't run on your robot. It needs a camera pipeline, an observation normaliser, an action de-normaliser, a safety wrapper, a replanning layer when it fails, and a way to log everything for later fine-tuning. <strong>OpenRAL is that infrastructure.</strong></p>
     <br/>
     <ul>
-      <li><strong>Typed runtime</strong> — eight well-defined layers connected by Pydantic v2 contracts. No magic globals, no hidden retries.</li>
-      <li><strong>rSkill format</strong> — Hub repos containing weights, a <code>rskill.yaml</code> manifest, quantisation hints, latency budgets, and reproducible <code>eval/</code>. Install like a model.</li>
-      <li><strong>Dual-system planning</strong> — a fast visuomotor policy (S1, 30–200 Hz) and a slow LLM planner (S2) emitting typed tool-calls. Replanning is bounded and explicit.</li>
-      <li><strong>Safety kernel</strong> — deny-by-default, Python proposes, C++ disposes. <code>ROSSafetyViolation</code> is never silently caught.</li>
     </ul>
   </div>
@@ -256,7 +256,7 @@ openral deploy run --config deployments/so100_pickplace.yaml</code></pre>
     <div class="arch-row">
       <span class="arch-num">0</span>
       <span class="arch-name">HAL</span>
-      <span class="arch-desc">Per-robot adapters — uniform connect / read_state / write_command Protocol</span>
       <span class="arch-tag tag-green">shipped</span>
     </div>
     <div class="arch-row">
@@ -274,7 +274,7 @@ openral deploy run --config deployments/so100_pickplace.yaml</code></pre>
     <div class="arch-row">
       <span class="arch-num">3</span>
       <span class="arch-name">rSkill (S1)</span>
-      <span class="arch-desc">Fast visuomotor policy — VLA, 30–200 Hz, async action chunks</span>
       <span class="arch-tag tag-green">shipped</span>
     </div>
     <div class="arch-row">
@@ -286,7 +286,7 @@ openral deploy run --config deployments/so100_pickplace.yaml</code></pre>
     <div class="arch-row">
       <span class="arch-num">5</span>
       <span class="arch-name">WAM</span>
-      <span class="arch-desc">World Action Model — mental simulation, failure anticipation</span>
       <span class="arch-tag tag-gray">planned</span>
     </div>
     <div class="arch-row">
@@ -309,11 +309,11 @@ openral deploy run --config deployments/so100_pickplace.yaml</code></pre>
     <table>
       <thead><tr><th>Type</th><th>What</th><th>Count</th></tr></thead>
       <tbody>
-        <tr><td><code>rskill-*</code> models</td><td>VLA policy rSkills — VLA weights + ROS-wrapped action skill + manifest + eval</td><td>20</td></tr>
-        <tr><td><code>rskill-*</code> models</td><td>ROS-action rSkills (<code>kind: ros_action</code>) — classical motion planners (MoveIt MoveGroup, Nav2) the reasoner dispatches like any other skill</td><td>4</td></tr>
-        <tr><td><code>rskill-*</code> models</td><td>Perception detector rSkills (<code>kind: detector</code>) — RT-DETR + OmDet-Turbo + LocateAnything → <code>ObjectsMetadata</code></td><td>5</td></tr>
-        <tr><td><code>rskill-*</code> models</td><td>Scene-understanding VLM rSkill (<code>kind: vlm</code>) — drives the reasoner's read-only <code>query_scene</code> tool</td><td>1</td></tr>
-        <tr><td><code>rskill-*</code> models</td><td>Reward / progress-monitor rSkill (<code>kind: reward</code>) — robotic reward model run parallel to a VLA; drives the reasoner's read-only <code>query_task_progress</code> tool</td><td>1</td></tr>
         <tr><td><code>dataset-*</code></td><td>LeRobotDataset v3 demonstration datasets</td><td>growing</td></tr>
       </tbody>
     </table>
@@ -417,7 +417,7 @@ openral deploy run --config deployments/so100_pickplace.yaml</code></pre>
   </div>
   <h2>ROS-action rSkills</h2>
-  <p><code>kind: ros_action</code> — classical motion planners packaged as rSkills, so the S2 reasoner dispatches a MoveIt or Nav2 motion exactly like a learned policy (this is the "VLA <em>and</em> classical controllers" thesis). No weights; they wrap a ROS 2 action server.</p>
   <div class="table-wrap">
     <table>
       <thead><tr><th>rSkill</th><th>Planner</th><th>What it does</th><th>License</th></tr></thead>
@@ -447,7 +447,7 @@ openral deploy run --config deployments/so100_pickplace.yaml</code></pre>
   </div>
   <h2>Perception detector rSkills</h2>
-  <p><code>kind: detector</code> — emit <code>ObjectsMetadata</code> (2D detections lifted to 3D via depth) rather than an <code>Action</code>. Plug directly into the <code>openral deploy</code> graph and fold into World State.</p>
   <div class="table-wrap">
     <table>
       <thead><tr><th>rSkill</th><th>Backbone</th><th>Notes</th><th>License</th></tr></thead>
@@ -469,7 +469,7 @@ openral deploy run --config deployments/so100_pickplace.yaml</code></pre>
         </tr>
         <tr>
           <td><code>rskill-omdet-turbo-locator</code></td>
-          <td>OmDet-Turbo (Swin-tiny)</td><td>On-demand <code>locate_in_view</code> — lightweight in-process "find X"</td>
           <td><span class="lic lic-green"></span>Apache-2.0</td>
         </tr>
         <tr>
@@ -482,7 +482,7 @@ openral deploy run --config deployments/so100_pickplace.yaml</code></pre>
   </div>
   <h2>Scene-understanding VLM rSkill</h2>
-  <p><code>kind: vlm</code> — a scene VLM that answers open-ended questions about the current view. It powers the S2 reasoner's <strong>read-only</strong> <code>query_scene</code> tool for task-progress / success verification ("did the grasp succeed?"); it holds no actuation authority (ADR-0047). Runs out-of-process in an NF4 sidecar served by <code>scene_vlm_node</code>.</p>
   <div class="table-wrap">
     <table>
       <thead><tr><th>rSkill</th><th>Backbone</th><th>Notes</th><th>License</th></tr></thead>
@@ -497,7 +497,7 @@ openral deploy run --config deployments/so100_pickplace.yaml</code></pre>
   </div>
   <h2>Reward-monitor rSkill</h2>
-  <p><code>kind: reward</code> — a robotic <strong>reward / progress-monitor</strong> model that runs <strong>in parallel with a VLA</strong> and scores the live rollout, emitting per-frame normalised progress (0–1) and success probability. It powers the S2 reasoner's <strong>read-only</strong> <code>query_task_progress</code> tool ("is the task succeeding right now?"); advisory, never on the control path (ADR-0057). Runs out-of-process in an NF4 sidecar served by <code>reward_monitor_node</code>; the pre-quantized checkpoint is meta-loaded directly as 4-bit.</p>
   <div class="table-wrap">
     <table>
       <thead><tr><th>rSkill</th><th>Backbone</th><th>Notes</th><th>License</th></tr></thead>
@@ -515,7 +515,7 @@ openral deploy run --config deployments/so100_pickplace.yaml</code></pre>
   <p style="font-size:12.5px; color: var(--mid);">
     <span class="lic lic-green"></span> Free for any use &nbsp;
     <span class="lic lic-yellow"></span> Research-permissive weights (check upstream terms) &nbsp;
-    <span class="lic lic-red"></span> Non-commercial — requires <code>OPENRAL_ACCEPT_NONCOMMERCIAL=1</code>
   </p>
   <!-- rSkill manifest format -->
@@ -545,7 +545,7 @@ quantization: null</code></pre>
   </div>
   <div class="robot-group">
-    <h3>Sim — HW bring-up in progress</h3>
     <div class="robot-chips">
       <span class="chip chip-blue">Franka Panda</span>
       <span class="chip chip-blue">UR5e</span>
@@ -580,7 +580,7 @@ openral benchmark report</code></pre>
   <!-- Observability -->
   <h2>Observability &amp; data flywheel</h2>
-  <p>Every skill execution is an OpenTelemetry span — weights revision pinned, camera frames captured, LLM prompts logged. Traces replay as LeRobotDataset v3 rows, closing the loop from deployment back to training data.</p>
   <pre><code>openral dashboard   # OTLP receiver at :4318, live trace viewer</code></pre>
   <!-- License -->

 <head>
   <meta charset="UTF-8" />
   <meta name="viewport" content="width=device-width, initial-scale=1.0" />
+  <title>OpenRAL: Runtime for VLA Robot Agents</title>
   <style>
     *, *::before, *::after { box-sizing: border-box; margin: 0; padding: 0; }
     <p>A VLA model alone can't run on your robot. It needs a camera pipeline, an observation normaliser, an action de-normaliser, a safety wrapper, a replanning layer when it fails, and a way to log everything for later fine-tuning. <strong>OpenRAL is that infrastructure.</strong></p>
     <br/>
     <ul>
+      <li><strong>Typed runtime</strong>: eight well-defined layers connected by Pydantic v2 contracts. No magic globals, no hidden retries.</li>
+      <li><strong>rSkill format</strong>: Hub repos containing weights, a <code>rskill.yaml</code> manifest, quantisation hints, latency budgets, and reproducible <code>eval/</code>. Install like a model.</li>
+      <li><strong>Dual-system planning</strong>: a fast visuomotor policy (S1, 30–200 Hz) and a slow LLM planner (S2) emitting typed tool-calls. Replanning is bounded and explicit.</li>
+      <li><strong>Safety kernel</strong>: deny-by-default, Python proposes, C++ disposes. <code>ROSSafetyViolation</code> is never silently caught.</li>
     </ul>
   </div>
     <div class="arch-row">
       <span class="arch-num">0</span>
       <span class="arch-name">HAL</span>
+      <span class="arch-desc">Per-robot adapters: uniform connect / read_state / write_command Protocol</span>
       <span class="arch-tag tag-green">shipped</span>
     </div>
     <div class="arch-row">
     <div class="arch-row">
       <span class="arch-num">3</span>
       <span class="arch-name">rSkill (S1)</span>
+      <span class="arch-desc">Fast visuomotor policy: VLA, 30–200 Hz, async action chunks</span>
       <span class="arch-tag tag-green">shipped</span>
     </div>
     <div class="arch-row">
     <div class="arch-row">
       <span class="arch-num">5</span>
       <span class="arch-name">WAM</span>
+      <span class="arch-desc">World Action Model: mental simulation, failure anticipation</span>
       <span class="arch-tag tag-gray">planned</span>
     </div>
     <div class="arch-row">
     <table>
       <thead><tr><th>Type</th><th>What</th><th>Count</th></tr></thead>
       <tbody>
+        <tr><td><code>rskill-*</code> models</td><td>VLA policy rSkills: VLA weights + ROS-wrapped action skill + manifest + eval</td><td>20</td></tr>
+        <tr><td><code>rskill-*</code> models</td><td>ROS-action rSkills (<code>kind: ros_action</code>): classical motion planners (MoveIt MoveGroup, Nav2) the reasoner dispatches like any other skill</td><td>4</td></tr>
+        <tr><td><code>rskill-*</code> models</td><td>Perception detector rSkills (<code>kind: detector</code>): RT-DETR + OmDet-Turbo + LocateAnything → <code>ObjectsMetadata</code></td><td>5</td></tr>
+        <tr><td><code>rskill-*</code> models</td><td>Scene-understanding VLM rSkill (<code>kind: vlm</code>): drives the reasoner's read-only <code>query_scene</code> tool</td><td>1</td></tr>
+        <tr><td><code>rskill-*</code> models</td><td>Reward / progress-monitor rSkill (<code>kind: reward</code>): robotic reward model run parallel to a VLA; drives the reasoner's read-only <code>query_task_progress</code> tool</td><td>1</td></tr>
         <tr><td><code>dataset-*</code></td><td>LeRobotDataset v3 demonstration datasets</td><td>growing</td></tr>
       </tbody>
     </table>
   </div>
   <h2>ROS-action rSkills</h2>
+  <p><code>kind: ros_action</code>. Classical motion planners packaged as rSkills, so the S2 reasoner dispatches a MoveIt or Nav2 motion exactly like a learned policy (this is the "VLA <em>and</em> classical controllers" thesis). No weights; they wrap a ROS 2 action server.</p>
   <div class="table-wrap">
     <table>
       <thead><tr><th>rSkill</th><th>Planner</th><th>What it does</th><th>License</th></tr></thead>
   </div>
   <h2>Perception detector rSkills</h2>
+  <p><code>kind: detector</code>. Emit <code>ObjectsMetadata</code> (2D detections lifted to 3D via depth) rather than an <code>Action</code>. Plug directly into the <code>openral deploy</code> graph and fold into World State.</p>
   <div class="table-wrap">
     <table>
       <thead><tr><th>rSkill</th><th>Backbone</th><th>Notes</th><th>License</th></tr></thead>
         </tr>
         <tr>
           <td><code>rskill-omdet-turbo-locator</code></td>
+          <td>OmDet-Turbo (Swin-tiny)</td><td>On-demand <code>locate_in_view</code>: lightweight in-process "find X"</td>
           <td><span class="lic lic-green"></span>Apache-2.0</td>
         </tr>
         <tr>
   </div>
   <h2>Scene-understanding VLM rSkill</h2>
+  <p><code>kind: vlm</code>. A scene VLM that answers open-ended questions about the current view. It powers the S2 reasoner's <strong>read-only</strong> <code>query_scene</code> tool for task-progress / success verification ("did the grasp succeed?"); it holds no actuation authority (ADR-0047). Runs out-of-process in an NF4 sidecar served by <code>scene_vlm_node</code>.</p>
   <div class="table-wrap">
     <table>
       <thead><tr><th>rSkill</th><th>Backbone</th><th>Notes</th><th>License</th></tr></thead>
   </div>
   <h2>Reward-monitor rSkill</h2>
+  <p><code>kind: reward</code>. A robotic <strong>reward / progress-monitor</strong> model that runs <strong>in parallel with a VLA</strong> and scores the live rollout, emitting per-frame normalised progress (0–1) and success probability. It powers the S2 reasoner's <strong>read-only</strong> <code>query_task_progress</code> tool ("is the task succeeding right now?"); advisory, never on the control path (ADR-0057). Runs out-of-process in an NF4 sidecar served by <code>reward_monitor_node</code>; the pre-quantized checkpoint is meta-loaded directly as 4-bit.</p>
   <div class="table-wrap">
     <table>
       <thead><tr><th>rSkill</th><th>Backbone</th><th>Notes</th><th>License</th></tr></thead>
   <p style="font-size:12.5px; color: var(--mid);">
     <span class="lic lic-green"></span> Free for any use &nbsp;
     <span class="lic lic-yellow"></span> Research-permissive weights (check upstream terms) &nbsp;
+    <span class="lic lic-red"></span> Non-commercial: requires <code>OPENRAL_ACCEPT_NONCOMMERCIAL=1</code>
   </p>
   <!-- rSkill manifest format -->
   </div>
   <div class="robot-group">
+    <h3>Sim: HW bring-up in progress</h3>
     <div class="robot-chips">
       <span class="chip chip-blue">Franka Panda</span>
       <span class="chip chip-blue">UR5e</span>
   <!-- Observability -->
   <h2>Observability &amp; data flywheel</h2>
+  <p>Every skill execution is an OpenTelemetry span: weights revision pinned, camera frames captured, LLM prompts logged. Traces replay as LeRobotDataset v3 rows, closing the loop from deployment back to training data.</p>
   <pre><code>openral dashboard   # OTLP receiver at :4318, live trace viewer</code></pre>
   <!-- License -->